jdk/src/java.base/share/classes/java/lang/module/ModuleFinder.java
author |
alanb |
|
Thu, 17 Mar 2016 19:04:16 +0000 |
changeset 36511 |
9d0388c6b336 |
child 37593 |
824750ada3d6 |
permissions |
-rw-r--r-- |
8142968: Module System implementation
Summary: Initial integration of JEP 200, JEP 260, JEP 261, and JEP 282
Reviewed-by: alanb, mchung, naoto, rriggs, psandoz, plevart, mullan, ascarpino, vinnie, prr, sherman, dfuchs, mhaupt
Contributed-by: alan.bateman@oracle.com, alex.buckley@oracle.com, jonathan.gibbons@oracle.com, karen.kinnear@oracle.com, mandy.chung@oracle.com, mark.reinhold@oracle.com, chris.hegarty@oracle.com, alexandr.scherbatiy@oracle.com, amy.lu@oracle.com, calvin.cheung@oracle.com, daniel.fuchs@oracle.com, erik.joelsson@oracle.com, harold.seigel@oracle.com, jaroslav.bachorik@oracle.com, jean-francois.denise@oracle.com, jan.lahoda@oracle.com, james.laskey@oracle.com, lois.foltan@oracle.com, miroslav.kos@oracle.com, huaming.li@oracle.com, sean.mullan@oracle.com, naoto.sato@oracle.com, masayoshi.okutsu@oracle.com, peter.levart@gmail.com, philip.race@oracle.com, claes.redestad@oracle.com, sergey.bylokhov@oracle.com, alexandre.iline@oracle.com, volker.simonis@gmail.com, staffan.larsen@oracle.com, stuart.marks@oracle.com, semyon.sadetsky@oracle.com, serguei.spitsyn@oracle.com, sundararajan.athijegannathan@oracle.com, valerie.peng@oracle.com, vincent.x.ryan@oracle.com, weijun.wang@oracle.com, yuri.nesterenko@oracle.com, yekaterina.kantserova@oracle.com, alexander.kulyakhtin@oracle.com, felix.yang@oracle.com, andrei.eremeev@oracle.com, frank.yuan@oracle.com, sergei.pikalev@oracle.com, sibabrata.sahoo@oracle.com, tiantian.du@oracle.com, sha.jiang@oracle.com
/*
* Copyright (c) 2014, 2016, Oracle and/or its affiliates. All rights reserved.
* DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
*
* This code is free software; you can redistribute it and/or modify it
* under the terms of the GNU General Public License version 2 only, as
* published by the Free Software Foundation. Oracle designates this
* particular file as subject to the "Classpath" exception as provided
* by Oracle in the LICENSE file that accompanied this code.
*
* This code is distributed in the hope that it will be useful, but WITHOUT
* ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
* FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
* version 2 for more details (a copy is included in the LICENSE file that
* accompanied this code).
*
* You should have received a copy of the GNU General Public License version
* 2 along with this work; if not, write to the Free Software Foundation,
* Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA.
*
* Please contact Oracle, 500 Oracle Parkway, Redwood Shores, CA 94065 USA
* or visit www.oracle.com if you need additional information or have any
* questions.
*/
package java.lang.module;
import java.io.File;
import java.io.FilePermission;
import java.nio.file.Files;
import java.nio.file.Path;
import java.nio.file.Paths;
import java.security.AccessController;
import java.security.Permission;
import java.security.PrivilegedAction;
import java.util.Collections;
import java.util.Objects;
import java.util.Optional;
import java.util.Set;
import java.util.stream.Collectors;
import java.util.stream.Stream;
/**
* A finder of modules. A {@code ModuleFinder} is used to find modules during
* <a href="Configuration.html#resolution">resolution</a> or
* <a href="Configuration.html#servicebinding">service binding</a>.
*
* <p> A {@code ModuleFinder} can only find one module with a given name. A
* {@code ModuleFinder} that finds modules in a sequence of directories, for
* example, will locate the first occurrence of a module of a given name and
* will ignore other modules of that name that appear in directories later in
* the sequence. </p>
*
* <p> Example usage: </p>
*
* <pre>{@code
* Path dir1, dir2, dir3;
*
* ModuleFinder finder = ModuleFinder.of(dir1, dir2, dir3);
*
* Optional<ModuleReference> omref = finder.find("jdk.foo");
* if (omref.isPresent()) { ... }
*
* }</pre>
*
* <p> The {@link #find(String) find} and {@link #findAll() findAll} methods
* defined here can fail for several reasons. These include include I/O errors,
* errors detected parsing a module descriptor ({@code module-info.class}), or
* in the case of {@code ModuleFinder} returned by {@link #of ModuleFinder.of},
* that two or more modules with the same name are found in a directory.
* When an error is detected then these methods throw {@link FindException
* FindException} with an appropriate {@link Throwable#getCause cause}.
* The behavior of a {@code ModuleFinder} after a {@code FindException} is
* thrown is undefined. For example, invoking {@code find} after an exception
* is thrown may or may not scan the same modules that lead to the exception.
* It is recommended that a module finder be discarded after an exception is
* thrown. </p>
*
* <p> A {@code ModuleFinder} is not required to be thread safe. </p>
*
* @since 9
*/
public interface ModuleFinder {
/**
* Finds a reference to a module of a given name.
*
* <p> A {@code ModuleFinder} provides a consistent view of the
* modules that it locates. If {@code find} is invoked several times to
* locate the same module (by name) then it will return the same result
* each time. If a module is located then it is guaranteed to be a member
* of the set of modules returned by the {@link #findAll() findAll}
* method. </p>
*
* @param name
* The name of the module to find
*
* @return A reference to a module with the given name or an empty
* {@code Optional} if not found
*
* @throws FindException
* If an error occurs finding the module
*
* @throws SecurityException
* If denied by the security manager
*/
Optional<ModuleReference> find(String name);
/**
* Returns the set of all module references that this finder can locate.
*
* <p> A {@code ModuleFinder} provides a consistent view of the modules
* that it locates. If {@link #findAll() findAll} is invoked several times
* then it will return the same (equals) result each time. For each {@code
* ModuleReference} element in the returned set then it is guaranteed that
* {@link #find find} will locate the {@code ModuleReference} if invoked
* to find that module. </p>
*
* @apiNote This is important to have for methods such as {@link
* Configuration#resolveRequiresAndUses resolveRequiresAndUses} that need
* to scan the module path to find modules that provide a specific service.
*
* @return The set of all module references that this finder locates
*
* @throws FindException
* If an error occurs finding all modules
*
* @throws SecurityException
* If denied by the security manager
*/
Set<ModuleReference> findAll();
/**
* Returns a module finder that locates the <em>system modules</em>. The
* system modules are typically linked into the Java run-time image.
* The module finder will always find {@code java.base}.
*
* <p> If there is a security manager set then its {@link
* SecurityManager#checkPermission(Permission) checkPermission} method is
* invoked to check that the caller has been granted {@link FilePermission}
* to recursively read the directory that is the value of the system
* property {@code java.home}. </p>
*
* @return A {@code ModuleFinder} that locates the system modules
*
* @throws SecurityException
* If denied by the security manager
*/
static ModuleFinder ofSystem() {
String home;
SecurityManager sm = System.getSecurityManager();
if (sm != null) {
PrivilegedAction<String> pa = () -> System.getProperty("java.home");
home = AccessController.doPrivileged(pa);
Permission p = new FilePermission(home + File.separator + "-", "read");
sm.checkPermission(p);
} else {
home = System.getProperty("java.home");
}
Path modules = Paths.get(home, "lib", "modules");
if (Files.isRegularFile(modules)) {
return new SystemModuleFinder();
} else {
Path mlib = Paths.get(home, "modules");
if (Files.isDirectory(mlib)) {
return of(mlib);
} else {
throw new InternalError("Unable to detect the run-time image");
}
}
}
/**
* Returns a module finder that locates modules on the file system by
* searching a sequence of directories and/or packaged modules.
*
* Each element in the given array is one of:
* <ol>
* <li><p> A path to a directory of modules.</p></li>
* <li><p> A path to the <em>top-level</em> directory of an
* <em>exploded module</em>. </p></li>
* <li><p> A path to a <em>packaged module</em>. </p></li>
* </ol>
*
* The module finder locates modules by searching each directory, exploded
* module, or packaged module in array index order. It finds the first
* occurrence of a module with a given name and ignores other modules of
* that name that appear later in the sequence.
*
* <p> If an element is a path to a directory of modules then each entry in
* the directory is a packaged module or the top-level directory of an
* exploded module. The module finder's {@link #find(String) find} or
* {@link #findAll() findAll} methods throw {@link FindException} if a
* directory containing more than one module with the same name is
* encountered. </p>
*
* <p> If an element in the array is a path to a directory, and that
* directory contains a file named {@code module-info.class}, then the
* directory is treated as an exploded module rather than a directory of
* modules. </p>
*
* <p> The module finder returned by this method supports modules that are
* packaged as JAR files. A JAR file with a {@code module-info.class} in
* the top-level directory of the JAR file is a modular JAR and is an
* <em>explicit module</em>. A JAR file that does not have a {@code
* module-info.class} in the top-level directory is an {@link
* ModuleDescriptor#isAutomatic automatic} module. The {@link
* ModuleDescriptor} for an automatic module is created as follows:
*
* <ul>
*
* <li><p> The module {@link ModuleDescriptor#name() name}, and {@link
* ModuleDescriptor#version() version} if applicable, is derived from
* the file name of the JAR file as follows: </p>
*
* <ul>
*
* <li><p> The {@code .jar} suffix is removed. </p></li>
*
* <li><p> If the name matches the regular expression {@code
* "-(\\d+(\\.|$))"} then the module name will be derived from the
* subsequence proceeding the hyphen of the first occurrence. The
* subsequence after the hyphen is parsed as a {@link
* ModuleDescriptor.Version} and ignored if it cannot be parsed as
* a {@code Version}. </p></li>
*
* <li><p> For the module name, then all non-alphanumeric
* characters ({@code [^A-Za-z0-9])} are replaced with a dot
* ({@code "."}), all repeating dots are replaced with one dot,
* and all leading and trailing dots are removed. </p></li>
*
* <li><p> As an example, a JAR file named {@code foo-bar.jar} will
* derive a module name {@code foo.bar} and no version. A JAR file
* named {@code foo-1.2.3-SNAPSHOT.jar} will derive a module name
* {@code foo} and {@code 1.2.3-SNAPSHOT} as the version. </p></li>
*
* </ul></li>
*
* <li><p> It {@link ModuleDescriptor#requires() requires} {@code
* java.base}. </p></li>
*
* <li><p> All entries in the JAR file with names ending with {@code
* .class} are assumed to be class files where the name corresponds
* to the fully qualified name of the class. The packages of all
* classes are {@link ModuleDescriptor#exports() exported}. </p></li>
*
* <li><p> The contents of all entries starting with {@code
* META-INF/services/} are assumed to be service configuration files
* (see {@link java.util.ServiceLoader}). The name of the file
* (that follows {@code META-INF/services/}) is assumed to be the
* fully-qualified binary name of a service type. The entries in the
* file are assumed to be the fully-qualified binary names of
* provider classes. </p></li>
*
* <li><p> If the JAR file has a {@code Main-Class} attribute in its
* main manifest then its value is the {@link
* ModuleDescriptor#mainClass() main class}. </p></li>
*
* </ul>
*
* <p> In addition to JAR files, an implementation may also support modules
* that are packaged in other implementation specific module formats. As
* with automatic modules, the contents of a packaged or exploded module
* may need to be <em>scanned</em> in order to determine the packages in
* the module. If a {@code .class} file that corresponds to a class in an
* unnamed package is encountered then {@code FindException} is thrown. </p>
*
* <p> Finders created by this method are lazy and do not eagerly check
* that the given file paths are directories or packaged modules.
* Consequently, the {@code find} or {@code findAll} methods will only
* fail if invoking these methods results in searching a directory or
* packaged module and an error is encountered. Paths to files that do not
* exist are ignored. </p>
*
* @param entries
* A possibly-empty array of paths to directories of modules
* or paths to packaged or exploded modules
*
* @return A {@code ModuleFinder} that locates modules on the file system
*/
static ModuleFinder of(Path... entries) {
return new ModulePath(entries);
}
/**
* Returns a module finder that is the equivalent to composing two
* module finders. The resulting finder will locate modules references
* using {@code first}; if not found then it will attempt to locate module
* references using {@code second}.
*
* <p> The {@link #findAll() findAll} method of the resulting module finder
* will locate all modules located by the first module finder. It will
* also locate all modules located by the second module finder that are not
* located by the first module finder. </p>
*
* @apiNote This method will eventually be changed to take a sequence of
* module finders.
*
* @param first
* The first module finder
* @param second
* The second module finder
*
* @return A {@code ModuleFinder} that composes two module finders
*/
static ModuleFinder compose(ModuleFinder first, ModuleFinder second) {
Objects.requireNonNull(first);
Objects.requireNonNull(second);
return new ModuleFinder() {
Set<ModuleReference> allModules;
@Override
public Optional<ModuleReference> find(String name) {
Optional<ModuleReference> om = first.find(name);
if (!om.isPresent())
om = second.find(name);
return om;
}
@Override
public Set<ModuleReference> findAll() {
if (allModules == null) {
allModules = Stream.concat(first.findAll().stream(),
second.findAll().stream())
.map(a -> a.descriptor().name())
.distinct()
.map(this::find)
.map(Optional::get)
.collect(Collectors.toSet());
}
return allModules;
}
};
}
/**
* Returns an empty module finder. The empty finder does not find any
* modules.
*
* @apiNote This is useful when using methods such as {@link
* Configuration#resolveRequires resolveRequires} where two finders are
* specified. An alternative is {@code ModuleFinder.of()}.
*
* @return A {@code ModuleFinder} that does not find any modules
*/
static ModuleFinder empty() {
// an alternative implementation of ModuleFinder.of()
return new ModuleFinder() {
@Override public Optional<ModuleReference> find(String name) {
Objects.requireNonNull(name);
return Optional.empty();
}
@Override public Set<ModuleReference> findAll() {
return Collections.emptySet();
}
};
}
}